Just wondering, since I have more than 10 1TB drives and a drobopro with 8x 2TB drives. Backing up is a very manual process. If a file changes and needs to be updated on the backup, I need to find out which backup drive it’s on and update it there. I know, it’d be ideal to just have another identical drobo set up as a backup for the first, but that’s a $2000 cost for convenience. Anyone heard of a backup program where you could select a large amount of data and have it spread it across many smaller drives automatically? I’m thinking something almost exactly like Toast’s disc spanning with optical media, but for hard drives.
Why don’t you make better use of that 10x1TB drives and form a RAID5/6 or 10 array and use any block-level, disk-based imaging solution w/ or w/o de-duplication to protect ur Pro? Not familarw/ Toast but will find out what it does. Tks for sharing!
LtCarter47, I don’t think you would be able to use a disk-spanning utility to do what you want.
Suppose you back up your Drobo and the data spans 4 drives: 3 are full and the 4th is half full. To make this backup, you just replace drives as they become full. So far, no problems.
Now suppose various files are deleted on your Drobo, and various other files are added. Overall there’s more data. You run the “sync” utility again. First, you’re prompted to connect “Drive 1.” Some files are deleted, but there’s not enough space to add the new ones. The utility simply prompts you to connect “Drive 4” which still has space, and it copies the new files. Then you’re prompted to connect Drive 2 to continue the backup, and so on. Drive 4 fills and the backup spans to Drive 5.
Your third backup starts getting complex. First you connect Drive 1, where some files are deleted, then Drive 4 where more files are deleted. Enough space exists now on Drive 1 to store some of the new files, so you connect Drive 1 again, but there’s not enough space for everything so you connect Drive 5 (which has space).
The backup continues to Drive 2, which then also has files added to both Drive 4 and Drive 5 from last time. Two more drive swaps. Then on to Drive 3…
After a few iterations of this you’ll become homicidal. “Homicidal” is fairly low on the Customer Satisfaction Scale at most companies, and tends to raise eyebrows among neighbors too.
What I suggest is that you divide up the data you wish to backup into 800-900GB chunks, either logically (photos, movies, VM images, humorous animated cat GIFs, the complete archives of alt.alien.vampire.flonk.flonk.flonk, etc) or roughly alphabetically by directory. Then on your Mac label each 1TB drive in sequence (“Backup 1” “Backup 2”…) This will give them unique mount points on the Mac. Finally, use a synchronization utility to backup each chunk whenever you like.
If you script the backups, or simply save preferences in the sync app so it will copy the data to the same target each time, the target volume will correspond to the disk sequence number you assigned when you started. If a drive gets full, simply shuffle some of the data to another drive and carry on.
You could either stuff each 1TB drive into its own enclosure, or use a “SATA dock” like those made by BlacX.
I personally use rsync and occasionally unison to sync my Drobo to external drives. I group directories by topic so that my photos are all together, my local CrashPlan data has its own disk, and so on. It’s hardly elegant, but it works.
Odd, I apparently missed all the replies to this thread and thought it had dropped off the front page unnoticed. Just got the email notification of the russian spam.
@Rambo, I like the idea, but it would probably mean a couple hundred dollars in drive enclosures and probably a big power strip to power them all.
Seems RDO is banned now, but he/she has some interesting thoughts. I was envisioning a backup utility that would keep a catalog of the drives involved in such a backup and intelligently plan the next “smart” backup so that the user would only need to insert, one time each, the disks to be modified in order to update the entire backup.
As for RDO’s idea on dividing the files up on the Drobo volume itself based on filesize and automating single drive backups from there, I think that’s a fantastic idea and I think it’s the way to go. With a program like SuperDuper! it could be totally painless aside from the initial organization and scheduling.
I now have a new project
Project well underway and going smoothly, this is a great system!
ArcServe does spanning backups. However, I think for a SOHO setup, using a online/near-line setup will be more efficient and cost-effective than spanning.
I’m not sure what you mean by SOHO or online/near-line, can you please explain?
Edit: Got it
SOHO = Small Office / Home Office. I think I get what near-line means from the wikipedia explanation. That mostly fits what I’m doing, except the robots, if only I had robots…